Estimates of Components of (Co)Variance by Restricted Maximum Likelihood for Carcass Traits in Rabbits Using an Animal Models

Abstract

AbstractGenetic and environmental (co)variance components were estimated for carcass trait records from 219 rabbits using univariate, restricted maximum likelihood (REML) with derivative and derivativefree REML (DFREML) methods. Derivative REML estimates of variance components using the GLMM (General Linear Mixed Models) package were for independent sire within sire breed, dam within sire, litter within dam, and residual random effects. Animal models used with DFREML included direct and maternal genetic effects with numerator additive genetic relationships, litter effects, and residual effects. Carcass trait heritabilities for direct genetic effects were generally consistent between corresponding REML methods. In animal models using DFREML that accounted for relationships, direct genetic variances were inflated when maternal genetic effects were ignored. Inclusion or exclusion of the litter effect had little impact on other variance components. Inclusion of covariance between direct and maternal genetic effects with numerator relationships resulted in relative shifts in direct and maternal genetic variances, but not in litter variances. Estimates of genetic correlations suggest antagonisms between direct and maternal genetic effects particularly for preslaughter, carcass, and total lean weight traits.